BILL GATES: Well, good morning, everyone, and I'll extend my welcome as well.

I get to kick things off by talking about the next 10 years of software, and some of the big breakthroughs that are going to be made, and how Microsoft has been investing in our research and our product groups and making sure that we drive these forward on our platforms, and together with our partners that we really seize the opportunity to have software change the world.

Now, this is what Microsoft has been about for over 30 years. But we would say the next 10 years are by far the most interesting. That is, the level of capability built into the software, the ability to help people work, to help them communicate, to help them create, help them understand what's going on in their business, to match buyers and sellers, to allow commerce to work in new ways, and to affect all the range of nonbusiness activities people are involved in is much stronger today than ever before.

I want to give you a sense of why we feel that way concretely: What are the enabling elements that have really come together? The vision from the beginning of the company was that we'd be focused on software, but we'd work with partners—partners who build hardware, who provide services, build applications—and that together we'd build these solutions for our customers. It was a bet on the microprocessor. At various times there's been milestones where there's been a bet on things like graphics interface, the rival of the Internet, the standard protocols around that; new platform opportunities, including what we did around the year 2000 with .NET; and taking the partners and moving them to that new level.

Many of the ways that we think about software, we have to get involved way before it becomes reality. A good example of that is the work we've done in TVs that you'll get an update on today, with what we called IPTV at the time—now called Mediaroom. We were looking at the Internet to deliver a high-quality video experience, to revolutionize the living room. And we've been working on that for over 10 years—a great example of where a willingness to invest in advance, apply research breakthroughs and build a platform that will allow for a whole new class of applications. TV will never be the same, and that's the same as all these digital software-empowered activities.

Now, why do we think this decade is so special? Well, it's partly the work we're doing and the incredible team around research and the products. But it's also the advances being driven by our partners to take trends and improvements and bring those to a dramatically higher level than ever before. First we have the constant improvement in the processor itself—that Moore's law, exponential increase in transistors, is not slowing down. There's a bit of a challenge in that those transistors will come to us without increasing the clock speed the way we've been able to just easily take advantage of for the last 20 years; that is, instead of doubling the clock like they've been able to, that piece will actually be going up at quite a modest rate. We may get up into the 10-gigahertz range, but not much higher than that, even, say, five or six years out. It will actually be multiple processors that allow for parallel execution will be a primary way that new silicon power is delivered. And there are lots of things in software development tools, in the operating system itself, that have to do with that parallel execution in making sure that you realize the full benefit of that powerful silicon. There's a lot of partnerships we have around this, including the one we have with Intel. But the silicon piece will continue to give us incredible opportunities.

The ubiquity of broadband actually is perhaps the most important thing here, because it's allowing us to think of computing in a different way. The early personal computer was a self-contained device. The graphics, the storage, the computation, was all in one device. When the user walked up to it, all of the state, all of the interaction, was with that single device. As you get broadband to be widely available, you can change that paradigm. You can think, OK, the storage doesn't all have to be in that location. It can be on servers run by that company or it can be servers on the Internet, or, as we often say, "in the cloud." And if you're working on multiple devices, you can move information back and forth between them in a far more user-centric way. If there's a display that's nearby that's bigger than the display on your device, you can connect up and put the information up on that display. If there's a computation resource that connects you to something a lot faster, then you can go across the network and take advantage of that. And so instead of the computer just being this one thing, it's actually all the resources on the network and the way those interact. And that's a fairly profound change. When we think about storage and the ability to make sure that it's backed up all the time and distributed in many locations so no disaster could cause you to lose that information, this Internet cloud-based approach is a fascinating way to be able to deliver on that. So it's far more user-centric, and you have a balance of computation—computation near the user, which is superior in terms of the responsiveness, low latency, always there, you don't have bandwidth limitations, or even if the network is unavailable, that's still there. But you also have the computation that's in datacenters or in these very large datacenters—we sometimes say mega datacenters that companies like Microsoft and others will create.

And so broadband enables the kind of browsing we have today, video to the Web; but it also lets us think about the computing paradigm in a new way.

The advance in the mobile devices is very exciting for us, because those have gone from being voice-only devices where software couldn't add much value, to now far-richer devices. If you look at what we're doing with Windows Mobile, what Apple is doing with the phone, it's about software innovation. The way you interact with the phone, the kind of data presentation you want there, some portion of your office capabilities brought onto that phone—the phone is part of the mix. In fact, in the course of the day, this user-centric approach will have you go from your car where you want to interact with information, to your TV set, to perhaps a PC at home, PC at work, phone—you may even have multiple of these pocket-size devices that you use at various times. And those will be very rich. But your interest, your information, your schedule—all of those will follow you as it's replicated through some sort of cloud-based capability. And so the opportunity for software on a wide range of devices is greater than ever before: unlimited storage, great-looking displays.

And the last one here, I think, is the one that's perhaps not as fully appreciated as some of the others: natural user interface. This is the idea that just typing and using the mouse is only one of the ways that we'll interact with these machines. The arrival of cheap cameras with great software, the ability with the pen to write on a surface, have that be recognized; the ability to recognize speech—those are going to come into the experience, and those are central to allowing computing to be pervasive. And in fact these things have way more impact than people expect when you describe them. In fact, I'll show the Surface here—of all the demos I've given in the many decades, this is the one that I've been the most surprised at how people respond to, simply because you're so used to the direct manipulation in that people can quickly envisage what kinds of things would become very easy for them to do in a group or alone that are very complex today.

So when Microsoft thinks about these 10 years and all these breakthroughs that we, because of our huge R&D investment and the great people we have, that we get to lead in, how do we organize that? How do we bring that together? Well, there's always a challenge in a product company that you want to have engineers who are very focused on the next two years, on that next release, and really thinking through practically the testing, the compatibility—all the issues that have to do with getting that next step forward done. But you also want engineers who are being wildly ambitious and saying, OK, as the storage moves and as the computation moves, how can this be different? That natural user interface becomes practical. And so that vision and the concrete engineering plans need to work together and often it's different individuals who have different skill sets in seeing those opportunities.

Well, we're organizing our engineering groups around these various communities—five communities and the platform itself—to really share our best ideas. And we go online and we use a SharePoint-type approach—it's a little bit like a wiki, but with some control over what the access is, where people contribute and share ideas. And then we take these visions, which are different than the product plans, and we just make sure the product plans are moving toward those visions, so we have the right investments, whether it's in pure research or incubation efforts in the product groups, to get to these advances.

So the five communities all rely on a common platform, and that platform is undergoing significant changes. It's always been a key asset for us, whether it's Windows or the Windows software running up on the server. And you'll hear a fair bit about this from Ray Ozzie—he's driving the revolutionary new platform that is service-centric, and so services plus software—software in the cloud, software in the datacenter, but then software on the rich client all working together in concert. So it's got—it brings the best capabilities of all those different tiers to work together.

Now, natural user interface—you're actually seeing a bit of this in Microsoft products and other products. If you think about why the Nintendo Wii has got such great approachability, it's that simple ability to think, OK, I'm swinging a bat. If you look at the iPhone, part of the appeal there is that touch capability that lets you get in and zoom things in and out and very quickly feel like you know exactly what you should do.

These natural user interface capabilities we've been investing in for a long, long time—speech, vision—with speech we recently did the Tellme acquisition that brought us even more capabilities there. So these are probably the most revolutionary things going on.

I mention Microsoft Surface and this was—we brought this out in late May, talked about it, showed it at the conference—and we talked about it with some partners, people like Harrah's, ITT, various companies that are going to put this into a retail or entertainment type environment. Now, that—it's not going to stop there. We want to take this, put it into homes and businesses. What is it? Well, this is a Windows Vista-based PC with a—some magic software and some cameras, so it can see when I touch the Surface—if it's turned on. Well, maybe we'll come back to that. It's more exciting when it actually does something—(laughter)—which right now it's not.

OK, the next—so I talked about five communities. The five communities include developers. Working closely with developers has always been a very big thing for Microsoft. You can go back to the earliest people who wrote basic applications or MS-DOS really getting ahead of its competitors because we worked with developers and just got more and more of these great applications. The vision of the broad community of developers and a very healthy software market were very key for all of our platform work.

To make things easier for developers, we need to do what we call adding the semantic level; that is, letting them express things in a higher-level form so they're not down in the details. If you think about a business process, you can describe that in English in a fairly succinct way. And yet, today, the code that that would require would be very, very large. And so how do you—why is that? Well, the answer is that we're not giving you the right semantic ability to describe it in terms of business-type activities. As you build declarative languages and proof engines into the platform, the difference between the simple English explanation and how the actual executable description gets done, you close that gap in a fairly dramatic way.

Now, this is very, very important, if we think about defining the software service and saying what sort of responsiveness we want, you want to do that at a very high semantic level. And then you want to rely on software to go out and find out what hardware, how many instances, where the storage should be, and make those decisions which require a lot of personnel and have been a very complicated thing for people to get involved in.

We need to make programming easy for information workers, for professionals, for people working in new domains doing things like robotics or ratification intelligence type work. Now, we have been raising this semantic level somewhat with things like Linq—it's a new set of language constructs that are very straightforward. We can go up to a level even beyond that with some of the new run-time capabilities. So we have our top people working in this area, and it will be part of the Office/SharePoint/Windows platform to let people write simpler applications.

The next community to talk about is the IT community itself. Our customers spend a lot of money organizing their machines and installing the software and wondering will it run reliably. There are a lot of error messages that can come out of the software, and so they have to be ready and have the expertise to understand all that stuff. Well, there are several techniques we can use to greatly reduce that complexity and to lower those costs. One thing is to take their datacenter and make it far more automatic. And of course we've learned a lot about doing that as we've done large scale datacenters ourselves for things like Web properties or even our internal needs. For example, our online site that delivers search results, there is no one who carries a beeper, because the software understands that if any of the hardware component fails, it knows how to redistribute that load. So it's completely fail-safe. There's multiple locations. That kind of capability can be provided to even fairly small-scale datacenters. And so people won't have to think about specific hardware mapping and specific performance type things; they just simply describe what they want the responsiveness to be. We've got virtual machine technology that's helping us here. We have some big advances in how we do security and how people describe security policy—where they think information should be available, where they think it should not. And so in fact the opportunity in this IT space is one of the biggest ones that we have.

Some of the traditional IT management approaches come down from the mainframe. And in fact it's just not the right paradigm. There's a radically different way of thinking about it. It uses some of that same modeling idea I talked about for developers. In fact, the gap between when you develop a program and when you want to run it, we need to eliminate that and so a lot of the high-level description gets done on the development side so that deployment becomes actually quite automatic. So this is a big area of investment. We talk about the products around the system center using our high-level run times as well.

The next audience is the information worker, and this is where the Office suite of products has been so incredibly successful. We've really defined what it's meant to create documents of all types and to have a user interface that's shared across the key applications.

Now we're taking that to a new level. SharePoint—you'll hear about the progress there—that's a big asset making it easy for people to write collaborative applications where you share things—very important. You'll hear about the investment in communication software, in business intelligence. The expectations of information workers need to be even higher than they are today, because software can allow them to see more patterns, to see things as they happen, to help them manage their time more effectively, to let them communicate in new ways.

A good example of that is all this stuff around telephony—bringing what we call asynchronous communications, things like e mail, together with synchronous, which is the phone, video-conferencing, screen-sharing, things of that nature. We have a big release of these products coming up in the fall, and part of that will be actually a hardware thing that we have called the Round Table. Some of you have seen this as we were developing it over the years, but it's now complete. This is what it looks like. It's just about a couple of thousand dollar product that you plug in. And what it's got here is it's got some cameras that are actually recognizing who's in the room, and it's got some microphones. It's far superior to just—even the audio part alone is way better than what came before. But it's actually video as well. And I think actually a lot of the people here will have situations where using this will be pretty interesting.

We've got a little video that will give you a little sense if you're a remote participant who's dialing in, connecting over the Internet to a meeting where there's a number of your colleagues in the room and you're trying to see what's going on with them, the video will give you a sense of what that experience is like. So let's --

(Video segment.)

BILL GATES: So, what you're seeing here is that the active speaker is shown in the upper left, and so the software notices that whoever is speaking and puts it up then is able to identify them.

Now, what it's done is, it's taken a 360-degree view, and it's brought that, the entire room, everybody is displayed across the top. So, at any time you can zoom out, just have your whole screen show the room, show the individual speaker, show the notes and slides, so you as a remote participant get a lot of control over what you want to see. In fact, we can record the meeting and have all these things available. So even if you're not connected up at the same time it's taking place, your ability to see it in rich ways, your ability to navigate, search through the information, search for the point where various people spoke up, that becomes very straightforward.

The information worker community, the idea of what will meetings be like in the future, what will information navigation be like, what will their desk look like, the ability to have a screen there will be pretty incredible. In fact, our original vision was a computer on every desk, and now we talk about even having a computer in every desk. Here with Surface that I believe is now responding, we can see why we're talking about that.

So what the camera does is it notices that you're touching it, and the software responds. And so anywhere you touch, the camera sees that's a hand. And so you can put lots of things on the surface here. You can put cards or dice or business cards. You can put down a phone. It's all recognized. And so it's easy to write different applications that run in this environment. We can take—here I'll take the photo application, which is a good one, and it just lets me play around with different photos that I might have. If I put the camera down, it'll connect over Bluetooth and let me actually bring in the photos that are there.

If I'm a user with an affinity card, it can recognize and read information there: you know, the manipulation of moving things around, setting the size. Here this is actually a video, so I can start that, even move it around while I play that. There's a lot of flexibility there, doing different things, looking at these different things. And the idea of organizing things or having multiple people share thoughts and say, "Oh, that looks pretty interesting"—very straightforward when you've got this type of surface. And no matter what you're doing, of course, you can put it away, switch and do something else pretty easily.

Here's an application that's actually being done with a partner. I think it'll show here this is the Sheraton. This is the idea that in a restaurant or a lobby, people who are interested in music then come up and see the different albums. When they put their affinity card down, it will understand their age or the things they might be interested in. I can take any one of these albums and select it and flip it over and look and say, "OK, what are the tracks?" take one of those tracks and put it into the queue in order to be played; take another album and take that, flip that over and put those in as well. Anything you want to do, it becomes fairly straightforward. And you're not actually having to learn anything here. Pretty quickly you'll get—you know, play around with the interface and get used to it.

Of course, if I have my music player, I can just put that down. It sees that that's a Zune and it'll connect up to it and be able—then I can take the songs that are there, look at those, play around with those. If I've got another Zune, I can put that on. People can exchange different music things; a lot of very simple scenarios that make sense.

Let me go back and just show one that shows kind of a new thing you might not even think about, but once you have this type of capability, it becomes very straightforward. Here what I've got is—I'll call it a video puzzle. It's just like a normal puzzle, except it's actually video. And so I've got these little cubes that I can put on. And if I put those on, what it does is it takes a video and splits it up into these different pieces. And then, as I move these around on the surface, it's moving the video along with it. And so what I have to do is figure out which of these things go together. I've given this demo about three times now, and I'm still not very good at doing this puzzle. It just shows it's actually fairly complicated to do. There, I got a few pieces together. I'm getting better all the time at this thing. Anyway, it's kind of fun and it just shows a very simple way that it's actually watching me move these things around and displaying the video on that.

And so you can imagine applications of quite a wide range. You know, if you're in your office and you want to call up the documents that relate to your meeting, your calendar, you want to navigate maps, things like that, your desk will have that built in. You can use your hands. You can write on the surface. This capability is very inexpensive, because those cameras are $10, $20 of hardware. But it's just software that sees what's going on.

You can think of this on horizontal surfaces, meeting rooms, the living room, or critical surfaces. What's been a white board will be a surface that you're displaying information on and then you go up and mark or change; the software will be able to play along with that. So it's starting to show how vision comes in as a fairly key element.

I've got two more communities. One is sort of business as a whole that we call rewiring the economy. And here's where you get business applications, the investments we've made there. We see the modeling approach, high-level descriptions of business processes, easy customization being important there, connecting in to have all of our horizontal office business intelligence immediately connected up to these applications. And this is also where we see online marketing, understanding the customer, personalizing things to them, the whole model of how advertising will change.

And, of course, that's just the beginning. The online advertising piece is a small piece of the overall advertising pie. But as TV viewing moves to be Internet-based, as even reading is more and more screen-based, the whole idea of targeting ads on having that interactive capability that we only have in that small piece today, that becomes completely a mainstream thing. And so that's why you're seeing us make investments in this area. We've got a lot of our top people thinking about this whole idea of how software facilitates matching buyers and sellers. In fact, we're announcing today an organizational step that'll help us do this even faster. It's the creation of what we call our Internet Services Research Center, and it's actually a group that combines our research capabilities and our product group capabilities. It'll be run by Harry Shum, who's being promoted to be a vice president. And he'll be working both for the research organization and actually for the group that does online capabilities, including our search-type activities.

The benefits for us of having this strong research group have been very important throughout our product line. But particularly in this dynamic area of advertising and search, they've really stepped up, and we've seen some quite advanced techniques that will allow us to show leadership in many aspects of this online activity; so driving the scale there, getting things going. There's a lot of vision about what will advertising look like over the next 10 years, because, as you shift virtually all of those dollars over, at least for young viewers, it'll be—the vast majority of dollars will be in the digital interactive environment. The need to have the right type of software there and help people move into that world is a huge opportunity for us moving forward.

The final community that we think about is everything you want to do at home. And, of course, it's expanded a lot how you think of computing helping out in your home activities. Photography today: Often, you know, people see digital as the best way to do that because you can organize, you can edit, you can share, you can tag. With music, more and more people are thinking this is the easiest way to have albums, share ideas of what good music is with other people. With video, this will be very mainstream, partly because the cameras will get so good, but also the wireless network will be such that you can immediately stream the video onto the storage in the home. And then, to back that up, it can be replicated out into the cloud. So really all the media types will move to be far superior to use them in a digital form.

I talked about the phone. The phone, like the personal computer itself, spans the world of business and the world of home activity. And it's fascinating in that sense, because you want to enable workers to do work at home or to organize home things like a trip or schedule changes while they're at work. And so they'll be interacting on the phone and on the personal computer. But essentially they'll have a different persona. The kind of things that they want to see that are important to them are different when you're in that context. And so making it easy to switch and see that sort of family view of your mail and calendar and go back to your work, and yet not have them completely separated, because, of course, it's one calendar, one device that you want to work with. That's a very interesting thing. And actually, one of these vision areas that we've written down has to do with making it simpler for people as they move between the multiple personas. Some people it's actually two, but for a lot of people, two are the key ones that they want to think about.

As I said, the phone is a key part of this. You know, think of the phone as being a remote control. Think of it as being a device you can do gaming on. The whole idea of the media stuff you carry around, the maps you want access to, your digital wall, all of those will be a variety of form factors.

And here's a case where we feel, just like with the personal computer itself, the benefit we get from having lots of great hardware partners who can do a variety of things is very helpful to us, but making sure that they're being very ambitious in terms of the interactivity, the high-end capabilities as they're building those phones.

The home will obviously continue to have TV coming into it. That's where the media room assets come in. The PC itself will be controlling both two-foot nearby and 10-foot further away type experiences. And so you want to move away from today's world, where we have many remote controls, to just either having one remote control or no remote control at all. In fact, as you get a camera in the living room, you can point and make gestures without even having a device of any type. There's a real question, do you need that single remote? Just getting down to one and having the full power of text on the screen that explains what's going on, explains your options, is a huge advance. But we think between vision and voice, there will be many cases that we won't need that at all.

Customizing to the individual is very straightforward when you can actually recognize their voice, see who they are, show them the shows or information that are of interest to them; create a ticker so that if they're watching a show, certain notifications of importance—somebody coming to the door, something changed, some upcoming scheduled things—that is individualized to whoever's there watching the TV.

We'd say that if you think about the future of TV, you're seeing a lot of that today in Xbox LIVE, where people are seeing which games their friends are playing, what they're up to, easy to get in touch with them. And an interesting milestone recently is when we connected up Xbox LIVE to the PC environment. And so your Messenger messages come across; your buddy list comes across. And we're not doing that with the phone as well. And so all of these properties and experiences, whether it's photos, calendar, messaging, all these things need to exist on the different devices. But you need the user interface to be somewhat tuned to the input technique and the screen size that you have at the time.

The way you want to look at Outlook on a full-size screen is very different than the way you want to look at it on your phone. Or if you're in the car, where driver distraction is a serious issue, the importance of using voice as part of that interaction becomes much greater. And so you need to bring that in as well. And so taking services and being able to make it easy to develop very rich front ends that are device-specific and yet personalized for that user, that's a big part of this environment. And we talked about it as a connected environment—connected media, connected entertainment, all of those things absolutely coming together.

So we have the five different communities and we've got the platform. The platform is the thing that pulls it all together. That's where you see the classic things about how is storage defined? How does it just show up where you want it in a magical way? How do you navigate that storage using searching and pegging and simple ways of organizing information that, if you're willing to mark something, then that gets preserved? It's that simple way of navigation.

Some things will be more automatic. For example, a location that you took a picture in will be automatically saved with that picture, just like the time and date is today. And so we'll be able to use that to create some really neat user interfaces where you can see a globe and look at the different pictures you've taken around the world and navigate in and see the pictures that other people have taken around the world. You'll also see something later today where we're taking pictures and we're analyzing them and realizing that two pictures are of a similar thing and being able to link, in a very automatic way, link and experience together. That's fun for personal photos. It's fun for photos of things like a store, where you want to create a new Web site experience.

Many of these experiences, we will move from being two-dimensional to having a three-dimensional experience. You're seeing this a little bit on the Web today. What you're going to see is that, for any Web site, the runtime that lets them create a 3-D experience will be easily available. And so they can control who comes in, what types of things go on. And if they want to connect up and use the same naming system or the same payment system as other Web sites, that's fine; like the 2-D Web site, it's just under their control. It's just the rich .NET environment that lets them build this type of 3-D navigation that they're interested in. We see that in commerce, where things like building a bookstore that's organized to you, or letting you see the real bookstore and walk around that. We see it for meetings. We see it for visualization. And we see it just for the pure entertainment world. That's part of the reason why we really like the investments we have in the rich runtimes that come out of our Xbox and Windows gaming-type environments, because many of those, as they get richer and richer, actually apply to domains other than entertainment.

Now, our focus as a company has always been software, continues to be software. That's where we see these big opportunities and seizing the things that I've talked about. That requires us to continue what we've been doing—building up the research centers all over the world; having, I think, the best connection between our peer research group and our product groups that any company has ever had. Many companies did great research but didn't manage to translate that into products in an agile way that they got the benefit of that research. Our experience, because we've put a lot of energy into hiring the right type of people, having the right processes, is that we're able to take those research investments and map those into significant competitive advantage.

Our development teams have really come together around this idea of the vision for software and the breathtaking things that we can do. So it's a lot of fun to see these different pieces coming together. You know, this really—this is the future of Microsoft, taking software to the next level, including a platform that lets partners build on that—hardware partners, software partners. All of them will be able to seize the same opportunities.

Thank you. (Applause.)

BILL GATES: With that, now I get to turn things over to Microsoft's CEO, Steve Ballmer. So, welcome, Steve.

END

Due to the varying sound quality and subject matter of tapes, the information in this transcript may contain inaccuracies.